• No results found

Has EMU had any Impact on the Degree of Wage Restraint?

N/A
N/A
Protected

Academic year: 2021

Share "Has EMU had any Impact on the Degree of Wage Restraint?"

Copied!
40
0
0

Loading.... (view fulltext now)

Full text

(1)

H

AS

EMU

HAD ANY

I

MPACT ON THE

D

EGREE OF

W

AGE

R

ESTRAINT

?

A

DAM

S.

P

OSEN

D

ANIEL

P

OPOV

G

OULD

CES

IFO

W

ORKING

P

APER

N

O

.

1783

C

ATEGORY

6:

M

ONETARY

P

OLICY AND

I

NTERNATIONAL

F

INANCE

A

UGUST

2006

An electronic version of the paper may be downloaded

from the SSRN website: www.SSRN.com

from the RePEc website: www.RePEc.org

(2)

CESifo Working Paper No. 1783

H

AS

EMU

HAD ANY

I

MPACT ON THE

D

EGREE OF

W

AGE

R

ESTRAINT

?

Abstract

We find in cross-sectional investigations that wage restraint is either unchanged or increased following EMU in the vast majority of countries. This contradicts the predictions of a widely-cited family of models of labor market bargaining. In those, Germany would have been expected to display the greatest decline in wage restraint post-EMU, and we find no indication of such a decline. The time-series evidence on Italy shows a significant increase in wage restraint after eurozone entry. This pattern is consistent with the models that emphasise the gains from monetary credibility. The eurozone increase in wage restraint is matched by the increase seen in the UK and Sweden after adopting inflation targeting, another means to credibility.

JEL Code: E58, E25, J58.

Keywords: EMU, wage bargaining, monetary credibility, productivity.

Adam S. Posen

Institute for International Economics 1750 Massachusetts Ave., NW

Washington, DC 20036-1903 USA

aposen@iie.com

Daniel Popov Gould

Institute for International Economics 1750 Massachusetts Ave., NW

Washington, DC 20036-1903 USA

July 2006

Posen’s work on this project was supported by a Houblon-Norman Fellowship at the Bank of England. We benefited from comments from our discussant John Driffill and from other participants in the Money, Macro & Finance Research Group and University Association for Contemporary European Studies Conference on ‘The Travails of the Eurozone,’ Edinburgh, March 24, 2006. The views expressed here and any errors are solely those of the authors, and not those of the Houblon-Norman Fund, the Bank of England, CESifo, or IIE. ©IIE, 2006.

(3)

Has EMU Had Any Impact on the Degree of Wage Restraint?

Adam S. Posen and Daniel Popov Gould

I. INTRODUCTION

The Lucas Critique notwithstanding, applied economic research has paid a great deal of attention in recent decades to the potential for changes in monetary regimes to induce lasting changes in

economic structures and behavior.1 In particular, given the key role of inflation expectations in

wage setting and the presumed endogeneity of practices such as indexation to the price environment, theorists have developed increasingly sophisticated models of the interaction

between central banks and labor market institutions.2

The creation of the euro presents a natural opportunity to investigate these models’ predictions. Eurozone member countries’ economies, where wage bargaining institutions and practices varied substantially, suddenly underwent a simultaneous shift in monetary regime to the European System of Central Banks (ESCB) and the Eurosystem. This shift would not necessarily have had the same effect on all eurozone economies—the pre-existing extent of unionization, degree of centralization and coordination in wage bargaining, relative size of the

economy in the monetary union, and so on could condition a given economy’s response.3 Given

this conditionality, theorists were generally building upon the insights of Calmfors and Driffill’s (1988) seminal paper on the interaction between centralization of wage bargaining and

macroeconomic performance.4 Interestingly, there was little agreement between the predictions

of European theorists and policymakers. Labor economists and political scientists whose

approach worked from wage bargaining institutions upward to macroeconomic outcomes tended to emphasize the risks to macroeconomic performance from Economic and Monetary Union (EMU) causing a mismatch between institutions, which might result in suboptimal coordination.

1 Lucas (1976) famously argued that known and lasting changes in economic policy would lead to offsetting private behavior such that policy changes would make little or no difference to economic outcomes.

2 See, among others, Bean (1998), Bayoumi and Sgherri (2004), Calmfors (1998), Cukierman and Lippi (1999, 2001), Duval and Elmeskov (2005), Gruener and Hefeker (1999), Hall and Franzese (1998), Iversen and Soskice (1998, 2000), Saint-Paul and Bentolila (2000), and Sibert and Sutherland (2000).

3 Calmfors (2001) and Cukierman and Lippi (2001) give useful albeit partial surveys of the factors involved. 4 These authors acknowledge their debt in turn to Olson (1982), who first outlined why the behavior of interest groups depends upon how “encompassing” their membership is.

(4)

On the other hand, almost all eurozone macroeconomists and central bankers felt that short-run inflation and output volatility would generally improve with the monetary regime shift, and many went further, suggesting that more credible policy would percolate downward and induce

structural reform in labor markets.5

Two difficulties have limited the success of empirical inquiries into this issue. The first, well-recognized, difficulty is that the number of country observations available to investigators is limited, and so therefore is the degree of freedom to distinguish robustly among competing

hypotheses.6 The second difficulty is that the vast majority of empirical studies tend to focus on

aggregate macroeconomic outcomes—normally inflation and/or unemployment—even though

the underlying theory usually generates hypotheses about real wage determination.7 The first

difficulty is largely unavoidable, though with time and with different cuts at the problem, some clarity may be achieved. The second difficulty, however, is unnecessary and may in fact

exacerbate the first problem. If the competing theories have different implications for real wages or functions thereof, it may be possible to distinguish between them by looking directly at the predictions for those variables and thus not burn degrees of freedom (and confuse the matter) by trying to control for shocks to inflation or unemployment or country-specific effects with respect to those variables.

This paper investigates the empirical implications of the various theories for EMU’s effect on wage restraint—the degree to which wage increases do or do not exceed productivity growth. In so doing, we attempt to avoid the seconddifficulty noted above. Wage restraint in some sense automatically controls for country-specific effects and shocks, beyond those directly accounted for in the observable labor and monetary institutions, because it is defined as a

response to a given country’s productivity performance, not a function of the level. It is also explicitly and frequently cited as a primary conscious concern of labor leadership, employers, and central bankers, who are presumed to be the actors in these models—and in the reality of

5 Posen (1998a, 1999) expressed an early American skepticism on both these counts. Duval and Elmeskov (2005) and Posen (2005) both give references to statements by euro proponents pre-EMU that the euro would strongly induce if not force structural change.

6 Blanchard (2005) is particularly articulate on the challenge presented to such cross-national datasets by the existence of a multiplicity of shocks. Calmfors (1993) and Driffill (2005) also note these limitations.

7 Richard Freeman raised this concern in initial comments on Calmfors and Driffill (1988). This is not true of all empirical investigations (e.g., Layard, Nickell, and Jackman [1991] estimate wage equations before turning to unemployment outcomes), but most focus on inflation or unemployment directly. Some of the models, such as those of Cukierman and Lippi (1999, 2001), map directly from real wages to aggregate outcomes.

(5)

wage bargaining and monetary policymaking. We also hope to partially alleviate the first difficulty, of inherently limited observations, by focusing on this dependent variable rather than inflation or unemployment, the determinants of which are perhaps too difficult to pin down. To our knowledge, no paper in this context has made wage restraint itself the focus of its empirical

work.8 In addition, taking advantage of the time that has passed since the launch of the euro, we

add to our cross-sectional work a comparison of time-series behavior in two critical countries, Germany and Italy, that differ markedly in what the theories predict would be the effect of EMU on wage restraint.

We find that wage restraint either is unchanged or has increased following EMU in the vast majority of countries, with no instances of significant declines in restraint. This finding contradicts the predictions of a widely cited family of models that rest on labor’s representatives in wage bargaining taking into account the external effect of their demands on aggregate

inflation. In particular, one would have expected Germany to display the greatest decline in wage restraint under these models, given that the relative importance of its major unions in the central bank’s calculus of inflationary pressures declined most with respect to both size and weighting by the central bank (given the shift from Bundesbank to the ECB), and we find no indication of such a decline. If anything, wage restraint in Germany increased post-EMU.

The overall shift in the countries examined toward greater wage restraint is consistent with the models that emphasize the gains to macroeconomic stability from monetary credibility, downplaying coordination or labor centralization issues. The time-series evidence on Italy, which shows a significant increase in wage restraint after eurozone entry, also supports this view, given that the restraint is mostly determined by the degree of monetary credibility (proxied by the Italian interest rate differential versus the lowest government bond rate in the eurozone). That said, the increase in wage restraint associated with increased monetary credibility in the eurozone is matched by that shown to be associated with the increase in credibility seen in the United Kingdom and Sweden after adoption of inflation targeting post-1992. This result emphasizes that the effect being seen is due to monetary regime changes and perhaps global pressures on labor

8 Of course, the concept of wage restraint and its importance is well established. Bruno and Sachs (1985) first brought in the modern concept of the wage gap and related it to institutions, and many of the papers cited in footnote 2 deal with real wage determination. Bean (1994, 2005), Blanchard (1991, 2000), Layard, Nickell, and Jackman (1991), and Nickell, Nunziata, and Ochel (2005) all consider the role of real wage rigidities in determining unemployment. Yet, the empirical linkage between wage restraint and changes in monetary policy regimes to our knowledge remains uninvestigated.

(6)

bargaining power and not due to EMU per se, political coordination issues, or international integration that accompanied EMU.

Section II further elaborates on our definition of wage restraint, its operationalization, and the distribution of outcomes in our sample. Section III goes through the implications of five major theories about the determinants of wage restraint for what should happen after EMU and whether or not non-eurozone advanced economies should see a similar effect. Section IV presents our cross-sectional analysis, looking at the extent of wage restraint before and after EMU in a sample of 19 economies (of which eleven are EMU members) and its determinants. Section V sets out our time-series analysis of the movements in wage restraint controlling for the business cycle in Germany and Italy and whether those dynamics changed after convergence in interest rates (our operationalization of monetary credibility) and the shift in monetary decision making from the Bundesbank to the ECB. Section VI considers the implications for two major theoretical approaches to the determinants of wage restraint and for policy, given what seems to be clear evidence in support of one and in apparent rejection of the other.

II. ROLE AND MEASUREMENT OF WAGE RESTRAINT

Central bankers and financial-market observers often refer to the degree of wage restraint in a country or in a given wage negotiation, by which they mean the degree to which increases in real wages are commensurate with increases in (labor) productivity. As argued by Bruno and Sachs (1985), the existence of a real wage gap—a persistent rise in real wages unmatched by

productivity—can explain stagflation in the 1970s. In a more recent example, the extensive and persistent high unemployment in the former East Germany is usually attributed to a lasting wage

bargain that overpriced eastern labor relative to its productivity.9 Others (e.g., Ball and Moffitt

2001, Blanchard and Philippon 2003) attribute part of the rise and in some countries fall of the nonaccelerating inflation rate of unemployment (NAIRU) to the lags with which wage setters recognized shifts in productivity growth. This attribution is of course loosely analogous to “classical” views of unemployment, where labor is overpriced relative to its returns, as opposed

9 See the summary and references in Posen (2006, chapter 6). The “overpricing” in turn can be attributed to the incentives for union insiders in then West Germany to prevent low-wage competition for their membership.

(7)

to more “Keynesian” views where nominal rigidities and insufficient demand are the root

cause.10

Assessment of wage restraint continues to play a significant role in the determination of monetary policy now that all central banks in advanced economies, including the ECB, are committed to forward-looking strategies for the stabilization of low inflation. The belief that wage increases “out of line with productivity” are potentially inflationary is widespread.

Consider the example of the German economy and Bundesbank behavior in the 1950–70 period, when economic growth was accompanied by union wage restraint, with real wages rising but less than the rate of growth in productivity—and the Bundesbank explicitly threatened to raise

interest rates should wage demands be “excessive” (Deutsche Bundesbank 1998, Streeck 1994 or

Siebert 2005, chapter 4).The decline in the US unemployment rate in the 1990s is widely

attributed in large measure to interaction between the outpacing of wage growth by productivity growth and the readiness of the Federal Reserve partly as a result to maintain low interest rates

even as past benchmarks for growth and unemployment were surpassed.11

It must be noted that wage restraint is not an entirely neutral concept distributionally. An increase in wages above the rate of productivity growth will embody some combination of

pass-through of inflationary expectations, of (mis)perception of the rate of productivity growth, and of

an increase in labor’s share of income relative to capital.12 Since ultimately factors of production

would be expected to earn their marginal products over the long term, this is less of an issue for multiyear averages than for any specific year’s wage settlement, but after some years of wage restraint, it could well be reasonable for labor’s share to catch up by growing above productivity

temporarily.13 As Caballero and Hammour (1996) and Blanchard and Philippon (2003) point

out, on average capital’s share of income has been rising along with unemployment in Europe

10 As Caballero and Hammour (1996) and Blanchard (2005) point out, though, even if a wage gap story can be used to explain much of the rise of European unemployment in the 1970s and early 1980s, it cannot be assigned a leading role in the persistence and, in some European countries, continued rise of unemployment in more recent years, precisely because there has been a period of relative wage restraint.

11 Blinder and Yellen (2001) articulate this position very well from the point of view of Federal Reserve decision makers.

12 As we are considering aggregate measures of wage restraint, we are abstracting from the bargaining over rents and quasi-rents between firms and unions, which are also a component of wage growth when considered at the sectoral or individual firm level.

13 In current monetary policymaking, the analogy would be made to a supply-shock that embodied a relative price shift. If the relative price of labor was going up due to structural reasons, and not simply as a response to broader price pressures, the central bank could accommodate the relative price shift by only gradually tightening in response to any inflationary effects and largely withholding any interest-rate response unless/until there were “second-round effects” of the wage increase on inflation expectations.

(8)

since the mid-1980s, suggesting that in recent years wage restraint has been ample and that the

real wage gap therefore is not the source of current unemployment.14 The relevant point for our

analysis is that we focus on year-to-year wage restraint because it is of declared and

demonstrated importance to monetary policy decisions, not because it is necessarily an optimal intermediate target for central banks or because it is welfare enhancing in and of itself.

Operationally, we define wage restraint as the difference between the rate of real wage growth and productivity growth for a given country in a given year. A negative (positive) observation indicates below (above) productivity wage growth. We use two measures of

productivity from the Organization for Economic Cooperation and Development (OECD). One is multifactor productivity (MFP) growth resulting as the residual from the OECD’s growth

accounting exercises; the other is growth in GDP per hour worked, calculated from national accounts data. The two measures are highly though not precisely correlated for multiyear averages, so we report all results below for both measures (where data are available). For consistency, we also rely on the OECD’s total compensation per employee measure from its

Economic Outlook as the source for our wage data. While we also undertook parallel

investigations with the OECD’s more narrowly defined “wage rate, business sector” series, we prefer total compensation because the data coverage is more complete and because the anecdotal evidence is that central banks pay more attention to changes in total compensation than in wages per se. Also, arguably there are wage negotiations where wage increases are kept low, but additional benefits with regard to pensions or the like are part of the package. To get real wage growth, we deflate each observation by the country-specific deflator (productivity growth is of course automatically in real terms).

Table 1 presents basic data on wage restraint for our full sample of 21 countries, comprising the 12 eurozone members, the three EU members outside the eurozone (Denmark, Sweden, and the United Kingdom), and six other advanced economies (Australia, Canada, Japan, Norway, Switzerland, and the United States). For three (Luxembourg, Norway, and

Switzerland), data are available only on GDP per hour growth and not on MFP growth. Since the focus of the paper is on the impact of EMU on wage restraint, we compute annual averages for

14 The outlier status of the US economy with regard to income and wealth inequality is not attributable to wage restraint either. As Dew-Becker and Gordon (2005) and Piketty and Saez (2006) recently demonstrated, the large increase in income going to the top 1 and 0.1 percent of earners in the United States over the recent decade is largely due to the extraordinary rise in executive pay cumulated over several years.

(9)

the pre-euro (1991–98) and post-euro (1999–2003 or 2004) periods, subject to data availability. The mean and median change in wage restraint in the sample is negative, meaning greater restraint in the post-euro period, for both measures of productivity. On productivity measured by MFP, the mean change is just short of being significantly different from zero at the 5 percent significance level (it is at the 7 percent significance level) and on the order of one percentage point (of the gap between productivity and wage growth) per year. Two countries, Greece and Sweden, show a significant increase in wage restraint under both measures, and two, Australia and the Netherlands, show a significant decrease in restraint under both measures, as can be seen more clearly in figure 1. Overall most EMU members showed no significant change in wage

restraint post-EMU.15

III. FIVE THEORIES OF THE DETERMINANTS OF WAGE RESTRAINT

At any given moment, both unions and individual workers are conducting wage negotiations with employers. Results of most negotiations are determined by idiosyncratic factors specific to sectors, firms, or individuals. Other negotiations are largely determined by automatic factors like cost of living allowances (COLAs), although their significance has declined since the 1970s. The overall macroeconomic environment, including productivity growth, however, also plays a role in wage negotiations, particularly since individual or industry productivity is often difficult to verify in real time. As bargaining becomes more collective and centralized, and especially when it takes place at a national level, such aggregate measures take on an even greater importance. Similarly, as central banks become more focused on maintaining low inflation (rather than reducing high inflation or pursuing other medium-term goals), the extent of wage pressures relative to productivity growth across the economy becomes a more salient issue. From these rather innocuous observations arise a number of theories about how labor and monetary institutions should influence wage restraint. Table 2 summarizes these theories with their empirical implication for restraint and whether they apply solely to eurozone members.

15 Greece clearly has by far the largest change in wage restraint. Dropping Greece from our sample would of course drive down the average change observed but also decrease the standard deviation of changes observed, so the number of (eurozone) countries showing a significant increase in wage restraint in fact rises in that subsample. Details available from the authors upon request.

(10)

The first set of theories builds directly on Calmfors and Driffill (1988) and concerns the extent to which unions take into account the inflationary impact of their wage bargains as a function of their membership. The more encompassing the membership (i.e., the greater the share of workers represented by the union), the more the union internalizes the cost of inflation

induced by wage pressures and thus the more likely it is to exercise wage restraint.16 In

particular, Cukierman and Lippi (1999, 2001) and Iversen and Soskice (1998, 2000) develop models of games between the ECB and unions as compared with the form of wage bargaining within countries before EMU. Both sets of papers predict that after EMU there will be a coordination problem. Unions that used to be large relative to their respective country’s total labor force, and whose bargaining their respective country’s central banks would have to take into account as a result when the latter was focused on national inflation rates, will after EMU become small(er) relative to the eurozonewide labor market and will not have to be taken into account by an ECB focused on eurozonewide inflation.

As a result, these models predict that wage restraint will decline after EMU. Unions’ incentives for wage restraint are reduced in two ways: one, “excessive” wage demands will have less effect on overall inflation so the cost to the unions’ members will be lower; two, and

probably more importantly, wage restraint from unions within one country will be less likely to induce monetary ease from the ECB—and therefore growth and employment increases—that benefits their members because the impact on eurozone-wide inflation will be smaller (countries where unions were already small or decentralized or absent would simply move further toward irrelevance for ECB monetary policymaking). So economies of large size before EMU where the central bank pursued an independent (nationally oriented as opposed to exchange-pegged) monetary policy should exhibit a significant decrease in wage restraint post-EMU—and this should be most marked in Germany, where not only was the economy the largest, while the Bundesbank most clearly took into account domestic wage developments when setting policy, but also the unions were large, and there was (is) nationwide wage bargaining. There is no reason to think that this shift should affect non-eurozone member countries.

16 In their famous U-shaped curve, Calmfors and Driffill (1988) suggest that extreme decentralization of wage bargaining will also lead to wage restraint because atomistic workers bargaining individually cannot drive up inflationary pressures. For purposes of considering the effects of EMU, the issue is whether economies on the “right” side of the hump with more concentrated wage bargaining move toward the suboptimal center where less internalization takes place, so we focus on that end here.

(11)

The second set of theories relates to the degree of international competition in product

markets and was formalized by Danthine and Hunt (1994). In this framework, unions have some

concern for the employment of their membership and recognize that employment will in part

depend upon the price competitiveness of their home country’s firms on world markets.17 If

wage increases outstrip productivity while other countries’ producers benefit from wage

restraint, the home country producers could lose market share, and the union members could lose jobs. As a result, the greater the exposure to international competition, the greater the wage restraint (Danthine and Hunt [1994] portray this as a shift in the Calmfors-Driffill curve). In the context of EMU, we can derive the prediction that to the extent that the introduction of the euro increased intra-eurozone trade, whether through increased transparency, lower transaction costs,

orother means, there should be greater wage restraint within the eurozone.18

A third set of theories comes more directly out of the political science tradition, though economists on the left in Europe sometimes support it. In this approach, the models of games between the ECB and labor unions become matters of outright bargaining between interest groups—where the ECB (like most central banks) is characterized as emphasizing inflation versus growth and employment objectives, while the labor representatives pursue the reverse (Garrett 1998, Hibbs 1987). The greater the political pull of the unions vis-à-vis the central bank, whether through threat of direct action because of union density and centralization or via the influence of elected representatives favoring union objectives, the lower the wage restraint because the central bank would be less willing (politically able) to “cut-off” growth in the economy. Absent the threat to tighten policy, the central bank would be unable to prevent a rise in labor share, which would mean real wage growth outpacing productivity. This approach would predict that after EMU, wage restraint would increase because the ECB would be less accountable to democratic control (given its insulation from national politicians), and there is no comparable Europe-wide labor institution to bargain on workers’ behalf. In particular, the countries where unions were more centralized and thus had greater political influence at home should see the greatest declines in wage restraint post-EMU.

17 Katzenstein (1984) first suggested this feedback effect in his study of small states in world markets. Such internalization of competitiveness concerns, however, is also a staple of policy discussions where there is tripartite bargaining. See, for example, Honohan and Lane’s (2002) depiction of the role of negotiated wage restraint in providing the conditions for the recent Irish miracle.

18 The discussion of the size of the increase in intra-European trade due to the adoption of the euro remains lively, with some very large estimates (e.g., Rose 2000) offered. See Baldwin (2005) and Frankel (2005) for a constructive debate over the accumulated empirical evidence.

(12)

A fourth set of theories has more to do with globalization and its impact on industrial democracies in general than with EMU per se. Given the effective rise in labor supply from emerging markets, which competes with production workers in advanced economies, and the increase in international capital mobility and institutions that make shifting of production to lower-cost sites easier, first-world workers face increasing pressure to remain competitive, if not decrease their unit labor costs. Add to this the more general trend toward deunionization in the major economies, or at least their private sectors, and the pressure for wage restraint should

increase (Dumont, Rayp, and Willemé 2006).19 This set of theories is in many ways parallel to

the second set of theories regarding openness to competition, but rather than emphasizing the change in incentives for given union structures and densities, this framework suggests a decline in those union densities. Thus the empirical prediction of this approach is that wage restraint should increase—both in and outside the eurozone—but primarily for the larger countries where labor was less subject to international competitive pressure in the past than in small countries that are already open.

The final set of theories of the determinants of wage restraint are those proposed by monetary economists and central bankers suggesting positive structural effects from EMU (e.g.,

European Monetary Institute 1998). In this framework, in economies where the central bank’s

commitment to price stability was less than credible, unions and workers had less incentive to take into account the costs of their own pursuit of inflationary wage settlements. On the one hand, their real wages were more likely to be eroded by increases in inflation, which would arise out of others’ wage and price expectations (and negotiations), so union negotiators would feel they had more at risk from wage restraint; on the other hand, the likelihood of short-term costs to employment from “excessive” wage settlements would be lower because the central bank would

be less credible in its threats to tighten policy should wage pressures rise.20 This is the converse

of the Bundesbank story behind the first set of theories discussed and as such is usually thought of as applying to Italy, for example, in the postwar period through the 1970s (or later). A rise in the credibility of central banks’ commitment to price stability should therefore induce greater wage restraint by reducing the fear that restraint will be self-defeating and increasing the fear

19 Whether the trend to deunionization is a result of these forces or is itself an independent cause, at least in part, is beyond the scope of this paper.

20 Obviously, we do not assume a stable trade-off between inflation and unemployment or the absence of costs to inflation, hence the mention of “short-term costs to employment,” since presumably central bank laxity would at some point induce real costs either from extra inflation and/or sharper tightening of policy.

(13)

that the central bank will not accommodate wage increases. This theory’s empirical prediction is that wage restraint should increase most for those countries that have the greatest increases in monetary credibility, whether through membership in EMU or through other means (such as the adoption of an inflation target).

IV. CROSS-SECTIONAL ANALYSIS OF POST-EMU CHANGES IN WAGE RESTRAINT

As discussed above, the limited number of observations available when considering these issues among the industrialized democracies encourages prudence in the use and interpretation of econometric analysis. Accordingly, when trying to sort out the impact of EMU on wage restraint, and the various theories of the determinants of wage restraint discussed in the preceding section, we stick to a simple approach. For those countries for which we have pre- and post-EMU

average wage restraint observations (a total of 18),21 we estimate ordinary least squares

regressions of the form:

∆WR = β0 + β1*eurodum + β2*(union variable) + β3*(union variable*eurodum)

+ β4*(∆monetary credibility) + β5*(∆monetary credibility*eurodum) + ε (1)

where “eurodum” denotes membership in the eurozone, “union variable” is a measure of union

density or centralization or coordination (taken from the literature), and “∆monetary credibility”

is a measure of the change in the credibility of the central bank’s commitment to price stability between the two periods.

We use different trade union variables as a robustness check of our results. The first

indicator is “trade union density,” obtained from the OECD’s Employment Outlook2004, which

utilized survey results to calculate this variable. We use the data on density in 1990 and 2000, labelling the variable “TUdense” in the estimation output. As a separate variable to control for trade union influence, we also use collective bargaining coverage, which is expressed as the fraction of the total labor force covered by collective bargaining. The OECD compiled the

(14)

collective bargaining coverage rates, which it took or estimated from several sources (including direct submissions from national governments). Wherever possible, coverage rates were adjusted for employees (particularly in the public sector) who do not have full rights to bargain, though

some public-sector workers obviously do (OECD’s Employment Outlook 2004, chapter 3). Once

again, we take values for 1990 and 2000 as pre- and post-EMU variables. Centralization and coordination of wage bargaining are the final two measures of trade union sway. These are constructed using survey data, also from the OECD, and are presented in the form of a cardinal scale from 1 to 5, increasing in half-point increments to indicate greater

centralization/coordination.

We proxy the change in monetary credibility by the difference between monthly long-term (10-year) government bond yields averaged for 1995m1 to 1997m12 and 1998m1 to

2000m12.22 A bigger difference indicates a larger decline in government bond rates and thus in

inflation expectations and in doubts about the central bank’s commitment to price stability. As shown in figure 2, there is a wide range of changes, with almost all economies in the sample seeing a minimum drop in bond rates of 150 basis points between the two periods considered, due presumably to global changes in inflation, the business cycle, and international arbitrage as the US and Japanese rates sank. Within Europe, as one might expect, Italy had the largest gain in credibility from EMU (more than 200 basis points above the eurozone average), with Portugal

and Spain gaining next most, and then Greece.23 Notably, however, Sweden and the United

Kingdom also had substantial drops in long bond rates, higher than the average eurozone member, over the period despite staying out of the eurozone.

Given the predictions of the various theories as outlined in table 2, the effects of wage bargaining structure and density, conditional (or not) on eurozone membership, and of monetary credibility should allow us to distinguish between them (if the data are willing). Table 3 presents our results from these cross-sectional (not panel) estimates. The three sets of estimates present a consistent set of results. An increase in monetary credibility has a significant increasing effect (at the 5 percent significance level) on wage restraint, and this result is not dependent on eurozone

22 The data are taken from the IMF’s International Financial Statistics. The first average ends in 1997m12 to allow for changes in bond yields in anticipation of EMU, and the horizons are shorter before and after than the wage restraint horizons to focus on the credibility impact of monetary regime shifts at the time.

23 Greece itself did not join the eurozone until January 2001, so the change seen here is assumed to have captured only part of EMU’s impact on the economy’s inflation expectations, with markets discounting until membership was sure.

(15)

membership (the (b) estimations) but is associated with all variation in monetary credibility in

the sample.24 The effect is economically meaningful as well with a 100 basis point drop in long

bond rates associated with a 0.42 percent increase in wage restraint (the differential between productivity growth and real wage growth in percent terms). Given an average for the sample of change (increase) in wage restraint of 0.68 percent, this is a substantial effect. Such a large effect associated with decreased long-term interest rates is consistent with the last of the theories discussed in the preceding section: that greater central bank credibility would induce greater wage restraint. Given the imprecise estimates of the constant terms and of most of the

coefficients on the other explanatory variables, changes in monetary credibility alone appear to

explain 40 percent of the cross-national variation in changes in wage restraint.25

Trade union density interacted with economy size (Tudense*gdp)shows up as significant

in the third regression, with a negative sign (it also has a negative coefficient significant at the 10 percent level in the first column, the other place it appears). The magnitudes are quite small, however, with coefficient estimates on the order of 0.00001. This would be weak evidence against the fourth set of theories discussed in the preceding section, since it implies that declining trade union density, conditional on being in a country large enough to have had some

independence of labor supply, decreases wage restraint. It seems consistent with Calmfors-Driffill’s underlying intuition that moving toward decentralization in the mid-range of

unionization would reduce incentives to restraint. The average economic size in our sample is

US$1,136.29 million (IMF’s World Economic Outlook 1999). This means that a fall in trade

union density from the sample average of 40 to 20 percent would lead to a nontrivial decline in wage restraint of 0.16 percentage point. Yet for the hump-shape argument, it is a problem that the sample mean of the density is 40 percent, since that would seem to put any declines in density to the right of the hump and therefore likely to lead to greater restraint.

Trade union density on its own, when not interacted with size of GDP, does not appear to be significant, just as the other measures of labor union organization and wage bargaining

24 Note that wage restraint is defined to be negative (it is wage growth minus productivity growth), so the more negative the number, the greater the restraint. This is why there is a positive coefficient on the credibility variable: cred1 is also negative, representing a fall in long-term rates from pre-EMU to post-EMU period.

25 Given the size of the increase in intra-eurozone trade, this result seems to be a particularly clear rejection either of the hypothesis that an increase in trade openness and competition should increase wage restraint, or of the

assumption that a removal of (intra-eurozone) currency variation should make trade more competitive. The relative clarity of this result may be in part a function of the more direct measurement of this explanatory variable than of changes in labor market institutions.

(16)

centralization do not, whether interacted with economic size or eurozone membership. This result is inconsistent with the third set of theories in section II: that declining political power or centralization of unions after EMU insulated monetary policymaking from dealing with labor would explain the observed increase in wage restraint. The extent of trade within the eurozone (EMUtrade2), presumably the exposure to international competition most directly affected by the launch of the euro, also does not show up as having a significant effect on wage restraint, as

opposed to the hypothesized positive coefficient.26 As with the monetary credibility hypothesis,

the determinants of wage restraint in the advanced countries appear on this data to be global (or by country) rather than associated with eurozone membership.

Particularly striking is the apparent rejection of the best formally developed theories of determination of wage restraint: those hypothesizing that a coordination problem would arise after EMU with the countries previously having unions that were large enough to internalize the costs of “excessive” wage demands showing a decline in wage restraint. As suggested in table 1, which shows the sample averages, the cross-sectional analysis in table 3 confirms that there is no association between the centralization or coordination of wage bargaining, whether conditional on size or not, and wage restraint for eurozone members—or for any countries in the sample. Since two distinct sets of models (Cukierman and Lippi 1999, 2001; Iversen and Soskice 1998, 2000) both make the strong prediction that wage restraint should have gone down after EMU, in

contrast to the other theories predicting conditional increases in restraint—andinstead wage

restraint went up on average, and significantly so even when taking account of the institutional factors underlying these models—it seems time to reconsider those models. Before doing so, however, we turn to time-series data to examine from another angle the empirical validity of the clear and contrasting predictions of the first (wage restraint down conditional on EMU and wage bargaining structure) and last (wage restraint up conditional on change in monetary credibility but not on EMU) theories from table 2.

26 Given the size of the increase in intra-eurozone trade, and the incidence of that expansion only for eurozone members, this result appears to be a particularly clear rejection of the trade competition increases restraint hypothesis. The relative clarity of this result may in part be a function of the more direct measurement of this explanatory variable than of the coding of labor market institutions, however.

(17)

V. TIME-SERIES ANALYSIS OF WAGE RESTRAINT IN GERMANY AND ITALY

We are now into the seventh year since the launch of the euro. It is feasible to undertake time-series analysis of even low-frequency data that span the periods before and after EMU, in order to look for EMU’s impact. With regard to wage restraint, the question is whether the adoption of the euro made any difference to year-by-year wage negotiations, given expected central bank reactions or nonreactions according to some theories. In the previous section we analyzed differences in multiyear averages for a set of 18 countries; in this section we turn to cyclical variation (or not) in wage restraint as a function of interest rates and structural factors for a pair of countries from 1980 to 2003. While we leave for future research the estimation of the

interaction between central bank reaction functions and wage equations, we hope to distinguish between the factors affecting wage restraint by carefully choosing the cases to consider.

Comparing the time-series behavior and determinants of wage restraint in Germany and Italy should allow us to see which effects of the euro are and are not evident. Germany was the economy with the de facto anchor currency of the pre-euro European Monetary System and had some of the largest unions with some of the most centralized and coordinated wage bargaining institutions in Europe. This combination of central bank independence (legally and in interest-rate setting) with centralized wage bargaining should have produced great incentives for wage restraint in Germany pre-EMU, according to the theories that emphasized incentives for union internalization of inflation costs. By the same token, the entry of Germany into the eurozone should have produced a marked drop in wage restraint—the German unions became notably smaller relative to the economic zone relevant for monetary policymaking, and monetary policymaking shifted away most clearly from a focus on German domestic inflation. In short, if the Cukierman-Lippi/Iversen-Soskice story in the spirit of Calmfors and Driffill should show up anywhere, it should be in a significant decline in wage restraint in Germany post-EMU. For fans of the monetary credibility story, there should either be no effect (assuming, as bond markets indeed seem to, that the ECB has just as credible a commitment to price stability as the Bundesbank did) or a slight decline in wage restraint.

For Italy, a different set of expectations is generated. As shown in figure 2, Italy enjoyed the largest credibility gain for its monetary policy commitment to low inflation upon admission

(18)

according to the theories emphasizing monetary credibility. In an economy where indexing was rife and inflation expectations were high and unanchored, there should have been little incentive for unions to exercise wage restraint—and little reason to think that the central bank would

tighten policy in response to excessive wage growth.27 After EMU, with a large gain in

counterinflationary credibility for Italian monetary policy (set by the ECB), Italian unions and wage bargainers should have shown a significant increase in wage restraint.

Even if the ECB were not setting policy on the basis of Italian wage developments, the eurozone more generally would be following a policy consistent with price stability, while the Italian economy would no longer be able to devalue or inflate at (political) will. If the monetary credibility story in the spirit of the postwar Bundesbank beliefs should show up anywhere, it should be in an increase in wage restraint in Italy post-EMU. Of course, according to the wage bargaining coordination problem theories, Italy as a large economy should be subject to a lesser version of the same phenomenon besetting German wage bargaining with the move into the eurozone and so should show no effect on, or a slight decrease in, wage restraint post-EMU.

To examine the determinants of wage restraint in these two critical country case studies, we look at annual data from 1980 to 2003 for compensation and productivity growth. In contrast

to the cross-sectional data on multiyear averages, here we utilize annual nominal compensation

growth (from OECD World Economic Outlook no. 78)along with contemporaneous GDP per

hour growth (from Groningen Growth and Development Centre Total Economic database as discussed in the previous section) to construct wage restraint. The switch from nominal to real wages is to take into account the money illusion and more broadly the difficulty for workers and unions in discerning real productivity growth in real time. The actual computations of real GDP per hour or of the residual from growth regressions that economists produce and we use above only appear with a lag usually of several months to actual events, whereas often wage

negotiations are on an annual or two- to three-year basis and are conducted in nominal terms. To examine the competing hypotheses, we estimate on German and Italian data separately regressions of the form:

27 Some commentators will insist that the Banca d’Italia did have significant counterinflationary credibility from the time of its “divorce” from the Italian Treasury or with the advent of later reforms. This begs credulity, given the revealed drop in long bond rates upon eurozone entry and the prior devaluations from the exchange rate mechanism (ERM)—let alone the desire of Banca d’Italia senior officials to gain eurozone entry.

(19)

WR = β0 + β1*Output Gap + β2*inflation expectations + β3*EMU dummy

+ β4*nominal central bank interest rate + β5*additional variables + ε (2)

where the additional variables include trade within the European Union, trade union density, and the spread between the country’s and Swiss long-term bond rates as a proxy for the nation’s central bank credibility. In the absence of time-series data on trade union density, we use the share of private-sector employment in the economy. Inflation expectations were obtained by

Chinn and Frankel (2003)—they take the average of month-to-month annual consumer price

index growth at 12-month leads and use it as an inflation expectation proxy. Our main interest is to see whether the EMU dummy is significant and negative, particularly for Germany, which would be consistent with the wage coordination story, or significant and positive, particularly for Italy, which would be consistent with the monetary credibility story.

Tables 4 and 5 present the results for Germany and Italy respectively. All regressions have 24 observations except those in column II of each table, where data limitations on the trade union density variable limit us to 13 observations. For Germany (table 4), we find the only factor significantly affecting wage restraint is the central bank instrument interest rate (which is

consistent with the Bundesbank wage restraint and deterrence story), with rises in that rate increasing restraint. Interestingly, the sole significance of this variable does not change after EMU, seemingly implying that German wage bargainers continue to keep their eye on the ECB response to their negotiations much as they did on the Bundesbank’s response. Surprisingly, even the German output gap and inflation expectations have no significant direct effects on wage restraint in Germany. There is no evidence in any of the estimates that a structural break

occurred around German economic unification in 1990–91, so we do not report separate results. This is not entirely surprising given the small share and separation of the eastern German labor market in overall German employment. Finally, there is evidence of a statistically significant (but not economically large) effect of the public budget deficit on wage bargaining (model VI), where a larger deficit increases wage restraint, perhaps in anticipation of either budget cutbacks or monetary response. This factor also accounts reasonably well for year-to-year variation in wage restraint in Germany in all estimates I-V even though the central bank interest rate is the sole significant explanator

(20)

For Italy (table 5), unemployment turns out to be a better measure than the output gap of the importance of the business cycle and has a consistently significant effect in the intuitive direction: an increase in unemployment increases wage restraint. Meanwhile EMU membership per se does not come in significantly for Italy, nor does the central bank interest rate itself— perhaps reflecting the de facto lack of independence of Italian monetary policy over the period. A direct measure of inflation expectations, however, is estimated to have a significant (5 or 10 percent level) and positive coefficient across most specifications, including ones where the EMU dummy is included, meaning that when there is a decrease in inflation expectations, wage

restraint increases. In a similar spirit, the spread between the Italian and Swiss long-term government bond rates has a significant positive coefficient; when the spread increases, consistent with a decline in Italian monetary credibility, wage restraint diminishes. That this shows up strongly in Italy is consistent with the predictions of the monetary credibility theory of wage restraint. Unlike in Germany, there is no evidence that budget deficits have any effect on wage restraint, but like in Germany the various measures of trade union structure and trade have

no discernable impact.28

VI. IMPLICATIONS FOR FUTURE RESEARCH

Enhanced monetary credibility, as proxied by the decrease in the long government bond rate after the launch of the euro, can explain a significant portion of the cross-sectional variation in the observed increase in wage restraint seen in European countries since 1999. The effect of monetary credibility on wage restraint is not limited to eurozone members, though; countries such as Sweden and the United Kingdom, which had a similarly measured gain in central bank commitment to price stability, also saw similar significant increases in wage restraint. This effect occurred completely independently of the wage bargaining institutions in the countries involved. In fact, in contradiction of the theories that suggested a coordination problem would emerge post-EMU between labor representatives and the ECB in large economies where bargaining was centralized and conducted with their national central banks prior to EMU, there is no evidence of a decline in wage restraint in those countries. A closer look at time-series evidence for Germany,

28 In both the German and Italian time series, one might expect the trade union coordination, centralization, and density variables to have limited explanatory power given their limited variation over the period.

(21)

where that hypothesized effect was supposed to be strongest, offers no support for the theories; time-series evidence for Italy on the variation of interest rate spreads over time, on the other hand, strongly supports the view that monetary credibility matters irrespective of wage bargaining arrangements.

As always in this literature, given the limited sample of countries involved and the limited data (both cross-sectional and time-series) available on institutional change, these results cannot be taken as dispositive. Yet, despite the data limitations, the results presented here are surprisingly robust and clear, perhaps benefiting from the focus on the narrowly defined

wage-restraint variable.Given the strength of the predictions of the Cukierman-Lippi (1999, 2001) and

Iversen-Soskice (1998, 2000) models that EMU should lead to a decline in wage restraint, particularly in large countries, the apparent rejection of those predictions should be taken seriously. The ECB has put a lower weight on individual countries’ cyclical and wage developments—particularly Germany’s—than the pre-EMU Bundesbank did when setting monetary policy (Posen and Popov Gould 2006; Hayo 2006), so the rejection is not because the ECB behaved contrary to expectations either in these models or more generally.

For future research, then, these results lead naturally to questions of what on the labor institution side was at work that coordination problems did not arise in wage bargaining post-EMU and, instead, wage restraint rose. When labor representatives appear on the basis of these results to be forward looking and concerned enough with macroeconomic conditions to respond to changes in counterinflationary credibility, it is somewhat surprising that the internalization dynamic for the effect of wage bargaining on inflation pressures does not carry through as well. Perhaps the labor representatives’ utility functions in the above models were simply

misspecified, with too little regard for employment effects and too much for the costs of

inflation. Shiller (1996) and survey work that followed established a healthy dislike for inflation among a wide range of the populace in many of the countries considered here, but that is not equivalent to establishing such a dislike among labor leaders, where anecdotal, political science, and historical evidence has tended to show labor as being far more concerned about output and employment than inflation (at least at low-to-moderate levels of inflation).

Another related possibility is that the importance of these labor and union institutions in wage bargaining behavior was overestimated by theorists, This could have occurred because actual coordination or centralization of labor bargaining in practice is fundamentally

(22)

mis-measured by the available coded classifications More moderately, while these institutions could

have had some sway, they also could have allowed for a great deal of variation in bargaining behavior over time, and those variations were not picked up in the available measures and thus drove down the estimates of the institutional codings’ impact our analyses. This scepticism may seem on the face of it unlikely, given the long emphasis on tripartite bargaining and corporatism in Europe, as well as the supposed recent successes of such mutual accommodation in Ireland, the Netherlands, and Sweden, and so the results here should not be used to impugn institutional factors altogether. Yet, it remains possible that the effect or effectiveness of these labor market institutions is endogenous to the political and economic forces in civil society and so produces the degree of wage restraint in keeping with the political pressures at any given time, largely

irrespective of apparent form.29

In any event, it may be necessary to go beyond investigating wage restraint (let alone unemployment or inflation outcomes) at the national level and consider sectoral differences in both wage bargaining structures and degree of wage restraint. Such an approach might not only better distinguish between these potential explanations for the absence of impact of wage bargaining structures on changes in wage restraint in the OECD in the last 15 years but also allow for more direct grappling with the alternative hypotheses advanced in section III,

particularly regarding the influence of globalization and competition on wage setting.30

For analysts of monetary policy, especially in the eurozone, at least one message is clear: The ECB has delivered wage restraint on the Bundesbank deterrence model where adoption of the euro led to credible declines in inflation expectations. This could be taken to indicate that concerns about establishing monetary toughness or the emergence of wage-push inflation pressures are unnecessary, especially since the adoption of inflation targeting in Sweden and the United Kingdom led to similar effects without any suggestion that they went through a similar proving process.

29 See Posen (1998b) for a general discussion of the endogeneity of institutional impact in political economy and macroeconomics.

(23)

REFERENCES

Baldwin, Richard. 2005. The Euro’s Trade Effects. Paper prepared for the ECB Workshop on

What Effects Is EMU Having on the Euro Area and Its Member Countries? Frankfurt, June.

Ball, Laurence, and Robert A. Moffitt. 2001. Productivity Growth and the Phillips Curve. In The

Roaring Nineties, ed., Alan Krueger and Robert Solow. New York: Russell Sage Foundation

Bean, Charles. 1994. European Unemployment: A Retrospective. European Economic Review

38: 523–34

Bean, Charles. 1998. The Interaction of Aggregate Demand Policies and Labour Market Reform.

Swedish Economic Policy Review 5, no. 2.

Bean, Charles. 2005. Comments on Blanchard. Economic Policy. 45:47-51.

Blanchard, Olivier Jean. 1991. Wage Bargaining and Unemployment Persistence. Journal of

Money, Credit, and Banking 23, no. 3 (August): 277–92.

Blanchard, Olivier. 2000. The Economics of Unemployment: Shocks, Institutions, and Interactions. Lionel Robbins Lecture, London School of Economics, London, October. Photocopy.

Blanchard, Olivier. 2005. European Unemployment: The Evolution of Facts and Ideas.

Economic Policy.45:6-47.

Blanchard, Olivier, and Philippon, Thomas. 2004. The Quality of Labour Relations and

Unemployment. MIT Department of Economics Working Paper 04-25 (June). Cambridge, MA:

Massachusetts Institute of Technology.

Blinder, Alan, and Janet Yellen. 2001. The Fabulous Decade. Washington: Century Foundation.

Bayoumi, Tamim, and Sgherri, Silvia. 2004. Monetary Magic? How the Fed Improved the

Flexibility of the U.S. Economy. IMF Working Paper WP/04/24 (February). Washington:

International Monetary Fund.

Bruno, Michael, and Jeffrey Sachs. 1985. The Economics of Worldwide Stagflation. Cambridge,

MA: Harvard University.

Caballero, Ricardo, and Mahmoud Hammour. 1998. Jobless Growth: Appropriability, Factor

Substitution, and Unemployment. Carnegie-Rochester Conference Series on Public Policy 48:

51–94.

Calmfors, Lars. 1993. Centralisation of Wage Bargaining and Macroeconomic Performance—A

Survey. OECD Economic Studies 21. Paris: Organization for Economic Cooperation and

(24)

Calmfors, Lars. 1998. Macroeconomic Policy, Wage Setting, and Employment—What

Difference Does the EMU Make? Oxford Review of Economic Policy 14.

Calmfors, Lars. 2001. Wages and Wage Bargaining Institutions in the EMU—A Survey of the Issues. Seminar Paper 690. Institute for International Economic Studies, University of

Stockholm, May.

Calmfors, Lars, and John Driffill. 1988. Bargaining Structure, Corporatism, and Macroeconomic

Performance. Economic Policy 6: 41–61.

Chinn, Menzie, and Jeffrey Frankel. 2003. The Euro Area and World Interest Rates. Working

Paper Series 1016. Santa Cruz, CA: Center for International Economics, University of California Santa Cruz.

Cukierman, Alex, and Francesco Lippi. 1999. Central Bank Independence, Centralisation of

Wage Bargaining, Inflation and Unemployment: Theory and Some Evidence. European

Economic Review 43, no. 7: 1395–434.

Cukierman, Alex, and Francesco Lippi. 2001. Labour Markets and Monetary Union: A Strategic

Analysis. Economic Journal 111: 541–65.

Danthine, Jean-Pierre, and Jennifer Hunt. 1994. Wage Bargaining Structure, Employment, and

Economic Integration. Economic Journal 104, no. 424: 528–41

Dew-Becker, Ian, and Robert Gordon. 2005. Where Did the Productivity Growth Go? NBER

Working Paper 11842 (December). Cambridge, MA: National Bureau of Economic Research. Driffill, John. 2005. The Centralisation of Wage Bargaining Revisited: What Have We Learned? London: Birkbeck College. Photocopy (November).

Deutsche Bundesbank. 1998. Fünfzig Jahre Deutsche Mark: Notenbank und Wärung in

Deutschland seit 1948. München.

Dumont, Michel, Glenn Rayp, and Peter Willemé. 2006. Does Internationalisation Affect Union

Bargaining Power? An empirical study for five EU-countries. Oxford Economic Papers

58:77-102.

Duval, Romain, and Jorgen Elmeskov. 2005. The Effects of EMU on Structural Reforms in

Labour and Product Markets. OECD Economics Department Working Paper No. 438 (July).

Paris: OECD.

Frankel, Jeffrey. 2005. Comments on Baldwin’s The Euro’s Trade Effects. Paper prepared for the ECB Workshop on What Effects Is EMU Having on the Euro Area and Its Member Countries? Frankfurt, June.

(25)

Garrett, Geoffrey. 1998. Parisan Politics in the Global Economy. Cambridge, UK: Cambridge University Press.

Groeningen Growth and Development Centre. 2006. Total Economic Data database. www.ggdc.net/index-dseries.html#top.

Gruener, Hans Peter, and Carsten Hefeker. 1999. How Will EMU Affect Inflation and

Unemployment in Europe? Scandinavian Journal of Economics 101, no. 1: 33–47

Hall, Paul, and Robert Franzese. 1998. Mixed Signals: Central Bank Independence, Coordinated

Wage-Bargaining, and European Monetary Union. International Organisation 52.

Hayo, Bernd. 2006. Is European monetary policy appropriate for the EMU member countries: a counterfactual analysis. Conference on The Travails of the Eurozone, Money, Macro, and Finance Research Group, Heriot-Watt University, March. Edinburgh.

Hibbs, Douglas. 1987. The Political Economy of Industrial Democracies. Cambridge, US: Harvard University Press.

Iversen, Torben, and David Soskice. 1998. Multiple Wage-Bargaining Systems in the Single

European Currency Area. Oxford Review of Economic Policy 14.

Iversen, Torben, and David Soskice. 2000. The Nonneutrality of Monetary Policy with Large

Price or Wage Setters. Quarterly Journal of Economics (February).

Layard, Richard, Stephen Nickell, and Richard Jackman. 1991. Unemployment, Macroeconomic

Performance and the Labour Market. Oxford: Oxford University Press.

Lucas, Robert. 1976. Econometric Policy Evaluation: A Critique. Carnegie-Rochester

Conference Series on Public Policy. 1:19-46

Nickell, Stephen, Luca Nunziata, and Wolfgang Ochel. 2005. Unemployment in the OECD

Since the 1960s: What Do We Know? Economic Journal 111, no. 500: 1–27.

OECD. 2004. Economic Outlook no. 78. Paris: Organization for Economic Cooperation and

Development.

Olson, Mancur. 1982. The Rise and Decline of Nations. Yale University Press.

Piketty, Thomas, and Emmanuel Saez. 2006. The evolution of top incomes. NBER Working Paper 11955 (January). Cambridge, MA: National Bureau of Economic Research.

Posen, Adam. 1998a. Central Bank Independence and Disinflationary Credibility: A Missing

(26)

Posen, Adam. 1998b. Do Better Institutions Make Better Policy? A Review Essay. International

Finance 1, no. 1: 173–205.

Posen, Adam. 1999. Why EMU is Irrelevant for the German Economy. Working Paper No. 99-05

(May). Washington: Institute for International Economics.

Posen, Adam, ed. 2005. The Euro at Five: Ready for a Global Role? Washington: Institute for

International Economics.

Posen, Adam. 2006 (forthcoming). Reform in a Rich Country: Germany. Washington: Institute

for International Economics.

Posen, Adam, and Daniel Popov Gould. 2006. ECB Monetary Policy and Divergences in the Eurozone. Institute for International Economics, Washington. Photocopy (June).

Rose, Andrew. 2000. One Money, One Market: Estimating the Effect of Common Currencies on

Trade. Economic Policy 30: 9–45.

Saint-Paul, Giles, and S. Bentolila. 2000. Will EMU Increase Euroclerosis? Discussion Paper

2423. London: Center for Economic Policy Research (April).

Shiller, Robert. 1996. Why Do People Dislike Inflation? NBER Working Paper No. W5539

(April). Cambridge, MA: National Bureau of Economic Research.

Sibert, Anne, and A. Sutherland. 2000. Monetary Union and Labour Market Reform. Journal of

International Economics 51.

Siebert, Horst. 2005. The German Economy: Beyond the Social Market. Princeton: Princeton

University Press.

Streeck, Wolfgang. 1994. Pay Restraint Without Income Policy: Institutionalised Monetarism

and Industrial Unionism in Germany. In The Return of Incomes Policy, ed., R. Dore, R. Boyer,

and Z. Marn. Pinter.

Visser, J. 2006. Union Member Statistics in 24 Countries. Monthly Labor Review.January,

(27)

VARIABLE LIST AND DATA SOURCES FOR CROSS-SECTION ANALYSIS

variable name variable label data source:

cntry country name MFPnom wage restraint: nominal

compensation growth minus MFP growth

GDPnom wage restraint: nominal compensation growth minus GDP

p/h growth OECD Economic Outlook (compensation) MFPreal wage restraint: real OECD Productivity DB (MFP, GDP p/h)

compensation growth minus MFP growth

GDPreal wage restraint: real

compensation growth minus GDP p/h growth

EMUtrade1 country's trade with eurozone

countries, avrg 1995-1998 IMF's DOTS database EMUtrade2 country's trade with eurozone

countries, avrg 1999-2004 cred1 credibility gain from euro

-pre-emptive convergence assumed govt long-term govt bond yield data from IMF's cred2 credibility gain from euro - NO IFS database

pre-emptive convergence assumed central1 centralization of bargaining

-1990-1994

central2 centralization of bargaining -1995-2000

coordin1 coordination index - 1990-1994 Driffill, John (2005) "The Centralization of wage coordin2 coordination index - 1995-2000 bargaining revisited. What have we learned?" collect1 collective bargaining coverage 1990

collect2 collective bargaining coverage 2000 Tudense Trade union density, 2000

gdp 1999 GDP, current prices, US$ IMF, World Economic Outlook Database billions, (WEO)

tuXXgdp interacted term - TU density in 1990 or 2000, GDP eurodum eurozone member dummy cred1euro interacted credibility with cred2euro euro dummy

central1gdp interacted centralization with gdp central2gdp

central1euro interacted centralization with euro d central2euro

coordin1euro interacted coordination with euro d coordin2euro

collect1gdp interacted collective bargaining cov. collect2gdp

(28)

VARIABLE LIST AND DATA SOURCES FOR TIME SERIES ANALYSIS: GERMANY, ITALY

variable name variable label Source of data

year time variable

restr_c_r wage restraint-contemporaneous,

real compensation growth

restr_l_r wage restraint-1 period lag of

productivity growth, real - Total Compensation data from OECD, Economic

compensation growth Outlook #78.

restr_c_n wage restraint-contemporaneous, - GDP p/h from: The Conference Board and

nominal compensation growth Groningen Growth and Development Centre,

restr_l_n wage restraint-1 period lagged Total Economy Database, January 2006

productivity growth, nominal compensation growth

gap output gap

nongovtempl share of total empl not in

public sector OECD, Economic Outlook #78

unempl unemployment rate

structdefchng % change in structural deficit

gdp gdp, billions, US$

TUden TU density

Visser, J. (2006) "Union member statistics in 24 countries", Monthly Labour Review

inflexpect inflation expectations IFS, 12m avrg of m-to-m CPI growth, 12m lead

r Buba/ECB money market interest

rate IMF, International Financial Statistics

spread Italian-Swiss LT govt bond

spread

trade_eu Italian trade with eurozone

countries-% GDP IMF, Direction of Trade database

trade_tot total Italian trade (X+M), % GDP

(29)

Table 1

Changes in average wage restraint

1991-98 vs. 1999-2004

Real Comp-MFP Real Comp-GDP p/h

Australia (2) 0.010* 0.012* Austria (1) -0.006 0.004* Belgium 0.004* 0.003 Canada -0.007 -0.002 Denmark 0.007* 0.003 Finland -0.007 0.002 France (2) -0.009 -0.006 Germany (3) -0.003 0.004* Greece -0.093* -0.096* Ireland -0.001 -0.001 Italy -0.007 -0.006 Japan (2) -0.015 -0.015 Luxembourg n.a. 0.004* Netherlands 0.011* 0.010* Norway n.a. 0.003* Portugal (1) -0.006 0.002 Spain (2) -0.012 -0.010 Sweden -0.025* -0.021* Switzerland n.a. -0.017* United Kingdom -0.022* -0.015 United States -0.004 -0.008 Avrg of EU12 -0.0117 -0.0086 Avrg of non-EU12 -0.0079 -0.0065 Mean -0.0102 -0.0068 Standard Error 0.0054 0.0046 Median -0.0065 -0.0013 Confid, Lvl. 95.0% 0.0113 0.0096 NOTES:

Differences between 1991-98 and 99-2004 averages of productivity growth and compensation growth

are subtracted from each other.

* signifies significantly different from the mean, at 5% (1) - MFP growth 1996-1999 average

(2) - MFP growth 1999-2002 average; (3) - 1992-1998 average

(30)

Table 2

Summary of Hypotheses on the Effect of EMU on Wage Restraint

Channel of transmission Effect on wage restraint Conditionality of effects Eurozone only? Relative size of external effects Decreases Larger on larger countries or those with indep. MP Yes Openness to international competition Increases Larger on those countries with high intra-EMU trade Yes Unions’ political bargaining power (not vs. firms) Increases

Larger for those countries with more centralisation Yes Decreased union density Increases

Larger for large countries where unions had more security from competition No Counter-inflationary credibility of central bank Increases

Larger for those countries who gain more credibility

References

Related documents

Method of Assessment • Written coursework submitted via email • dissertation • No use of web-based online assessment tool • Learning activities • Formative

Microscopic images of the igneous rocks collected for studied area (a) plagioclase, olivine, microlite plagioclase phenocrysts in the basaltic andesite olivine

characterize Begomovirus taken from yard long bean samples in Java based on specific genome property of common region. Three main activities were conducted: i) sample

As it is meant to harm the group or a community, by intimidating other people in the group, by fragmenting or destroying the community ties (with fear, shame, stigmatization

PCR followed by sequencing of DNA extracted from blood or spleen of the infected animals revealed 99–100% identity of the amplified part of the 18S rRNA gene with GenBank

• Storage – If unopened, optimum shelf life is up to 30 years. If opened should be used within

In this research, a new gait energy-based feature, backfilled gait energy image (BGEI), that can be constructed from both side view silhouettes and frontal depth images, is

Since the seminal workshop, 19 cluster events have been organized, and special issues and books have been organized on accessibility modelling ( Mart´ın, van Wee 2011 ,